09:15 〜 09:30
[HDS08-02] Monitoring landslide frequency and vegetation recovery using satellite time series data
キーワード：Landslide frequency, vegetation recovery, Google Earth Engine (GEE), NDVI time series, Taiwan
The prevalence of widespread landslides in subtropical countries, such as Taiwan, resulting from earthquakes, typhoons, and heavy rainfall poses significant risks to both human lives and the country's economy. Analyzing such natural disasters is critical for estimating the risks they pose because future incidents are likely to occur under conditions similar to previous ones. Hence analyzing time patterns and regional occurrence of triggering events will aid in the comprehensive study of landslides. Remote sensing technique provides geographic coverage and frequent acquisitions even in dangerous and unreachable regions and aid in understanding landslide factors. This study uses normalized difference vegetation index (NDVI) time series data from Landsat 5 and Landsat 8 in the Google Earth Engine (GEE) environment to present an automated technique for long-term landslide identification in Taiwan to get over geographical and temporal limitations and thoroughly evaluate the incidence of landslides. Five major landslide-prone zones of Taiwan are examined in this study to understand the landslide occurrence frequency and vegetation recovery duration. The landslide frequency (LSF) exhibits a diverse distribution, ranging from 1 to 6, signifying varied occurrences of landslides in Taiwan. Across Taiwan, distinct temporal patterns unfold, with certain regions marked by recurring landslides impeding vegetation regrowth, while others showcase swift recovery. This study contributes to an enhanced comprehension of landslide dynamics through the application of automated detection and analysis methods on a comprehensive regional scale.